Automatic Handwriting Generation

Slides:



Advertisements
Similar presentations
A brief review of non-neural-network approaches to deep learning
Advertisements

Lecture 14 – Neural Networks
Distributed Representations of Sentences and Documents
Attention Deficit Hyperactivity Disorder (ADHD) Student Classification Using Genetic Algorithm and Artificial Neural Network S. Yenaeng 1, S. Saelee 2.
Convolutional Neural Networks for Image Processing with Applications in Mobile Robotics By, Sruthi Moola.
Handwriting Copybook Style Analysis Of Pseudo-Online Data Student and Faculty Research Day Mary L. Manfredi, Dr. Sung-Hyuk Cha, Dr. Charles Tappert, Dr.
Richard Socher Cliff Chiung-Yu Lin Andrew Y. Ng Christopher D. Manning
COMPARISON OF IMAGE ANALYSIS FOR THAI HANDWRITTEN CHARACTER RECOGNITION Olarik Surinta, chatklaw Jareanpon Department of Management Information System.
Dr. Z. R. Ghassabi Spring 2015 Deep learning for Human action Recognition 1.
Handwritten Recognition with Neural Network Chatklaw Jareanpon, Olarik Surinta Mahasarakham University.
Deep Learning Overview Sources: workshop-tutorial-final.pdf
By: Shane Serafin.  What is handwriting recognition  History  Different types  Uses  Advantages  Disadvantages  Conclusion  Questions  Sources.
Speech Recognition through Neural Networks By Mohammad Usman Afzal Mohammad Waseem.
Bassem Makni SML 16 Click to add text 1 Deep Learning of RDF rules Semantic Machine Learning.
Tofik AliPartha Pratim Roy Department of Computer Science and Engineering Indian Institute of Technology Roorkee CVIP-WM 2017 Paper ID 172 Word Spotting.
Course Outline (6 Weeks) for Professor K.H Wong
Big data classification using neural network
Hybrid Deep Learning for Reflectance Confocal Microscopy Skin Images
Unsupervised Learning of Video Representations using LSTMs
Learning to Compare Image Patches via Convolutional Neural Networks
Analysis of Sparse Convolutional Neural Networks
Faster R-CNN – Concepts
CS 388: Natural Language Processing: LSTM Recurrent Neural Networks
The Relationship between Deep Learning and Brain Function
Machine Learning for Big Data
Deep Learning Amin Sobhani.
Randomness in Neural Networks
Convolutional Neural Fabrics by Shreyas Saxena, Jakob Verbeek
Recurrent Neural Networks for Natural Language Processing
COMP24111: Machine Learning and Optimisation
Matt Gormley Lecture 16 October 24, 2016
Combining CNN with RNN for scene labeling (segmentation)
Intelligent Information System Lab
Intro to NLP and Deep Learning
Synthesis of X-ray Projections via Deep Learning
Natural Language Processing of Knee MRI Reports
Final Year Project Presentation --- Magic Paint Face
Shunyuan Zhang Nikhil Malik
CS6890 Deep Learning Weizhen Cai
Master’s Thesis defense Ming Du Advisor: Dr. Yi Shang
By: Kevin Yu Ph.D. in Computer Engineering
Layer-wise Performance Bottleneck Analysis of Deep Neural Networks
Bird-species Recognition Using Convolutional Neural Network
A critical review of RNN for sequence learning Zachary C
Introduction to Neural Networks
Grid Long Short-Term Memory
Image Captions With Deep Learning Yulia Kogan & Ron Shiff
CSC 578 Neural Networks and Deep Learning
Introduction to Deep Learning with Keras
Final Presentation: Neural Network Doc Summarization
March 2017 Project: IEEE P Working Group for Wireless Personal Area Networks (WPANs) Submission Title: Deep Leaning Method for OWC Date Submitted:
network of simple neuron-like computing elements
Machine Learning 101 Intro to AI, ML, Deep Learning
Creating Data Representations
The Big Health Data–Intelligent Machine Paradox
On Convolutional Neural Network
Zip Codes and Neural Networks: Machine Learning for
实习生汇报 ——北邮 张安迪.
Presentation By: Eryk Helenowski PURE Mentor: Vincent Bindschaedler
Deep Learning Authors: Yann LeCun, Yoshua Bengio, Geoffrey Hinton
Department of Computer Science Ben-Gurion University of the Negev
Object Detection Implementations
Bidirectional LSTM-CRF Models for Sequence Tagging
CSC 578 Neural Networks and Deep Learning
Listen Attend and Spell – a brief introduction
Food Dish Classification using Convolutional Neural Networks
Huawei CBG AI Challenges
An introduction to Machine Learning (ML)
Shengcong Chen, Changxing Ding, Minfeng Liu 2018
Presentation transcript:

Automatic Handwriting Generation Presented by Ms.S.T.shenbagavalli, AP/CSE

Deep Learning Deep Learning is changing the way we look at technologies. There is a lot of excitement around Artificial Intelligence (AI) along with its branches namely Machine Learning (ML) and Deep Learning at the moment.

Traditional handwriting Generation system Handwriting recognition - classifying each handwritten document by its writer is a challenging problem due to huge variation in individual writing styles. The traditional approach to solving this would be to extract language dependent features like curvature of different letters, spacing b/w letters etc. and then use a classifier like SVM to distinguish between writers. 

Hand writing recognition using deep learning is a very powerful technique for several reasons: It automatically identifies deep powerful features Our approach of feeding in random patches makes the model text independent High prediction accuracy makes it possible to use this in practical applications

Handwritten Text Recognition (HTR)

INTRODUCTION This is a task where given a corpus of handwriting examples, generate new handwriting for a given word or phrase. The handwriting is provided as a sequence of coordinates used by a pen when the handwriting samples were created. From this corpus, the relationship between the pen movement and the letters is learned and new examples can be generated ad hoc.

Step 1:Converting into digital text Offline Handwritten Text Recognition (HTR) systems transcribe text contained in scanned images into digital text

Step-2:Building the neural network Layer Build a Neural Network (NN) which is trained on word-images from the IAM dataset. As the input layer (and therefore also all the other layers) can be kept small for word-images, NN-training is feasible on the CPU (of course, a GPU would be better). 

Model Overview- HTR system

It consists of convolutional NN (CNN) layers, Recurrent NN (RNN) layers and a final Connectionist Temporal Classification (CTC) layer.  We can also view the NN in a more formal way as a function which maps an image (or matrix) M of size W×H to a character sequence (c1, c2, …) with a length between 0 and L.

Operations CNN The input image is fed into the CNN layers. These layers are trained to extract relevant features from the image. Each layer consists of three operation. First, the convolution operation, which applies a filter kernel of size 5×5 in the first two layers and 3×3 in the last three layers to the input. The non-linear RELU function is applied. Finally, a pooling layer summarizes image regions and outputs a downsized version of the input. 

Operations RNN The feature sequence contains 256 features per time-step, the RNN propagates relevant information through this sequence. The popular Long Short-Term Memory (LSTM) implementation of RNNs is used, as it is able to propagate information through longer distances and provides more robust training-characteristics than vanilla RNN.

Operations CTC While training the NN, the CTC is given the RNN output matrix and the ground truth text and it computes the loss value. While inferring, the CTC is only given the matrix and it decodes it into the final text. 

Implementation using TF The implementation consists of 4 modules: SamplePreprocessor.py: prepares the images from the IAM dataset for the NN DataLoader.py: reads samples, puts them into batches and provides an iterator-interface to go through the data Model.py: creates the model as described above, loads and saves models, manages the TF sessions and provides an interface for training and inference main.py: puts all previously mentioned modules together

Conclusion The NN consists of 5 CNN and 2 RNN layers and outputs a character-probability matrix. This matrix is either used for CTC loss calculation or for CTC decoding. An implementation using TF is provided and some important parts of the code were presented. Finally, hints to improve the recognition accuracy were given.